DocumentCode :
2773804
Title :
A proposal for human action classification based on motion analysis and artificial neural networks
Author :
Rocha, Thiago Da ; De Barros Vidal, Flavio ; Romariz, Alexandre Ricardo Soares
Author_Institution :
Dept. of Electr. Eng., Univ. of Brasilia, Brasilia, Brazil
fYear :
2012
fDate :
10-15 June 2012
Firstpage :
1
Lastpage :
6
Abstract :
This paper describes the development and application of a method for human action recognition from motion analysis in a sequence of images using an artificial neural network. The proposed method is based on two stages: Computer Vision and Computational Intelligence. The Computer Vision stage is a combination of two motion analysis techniques: Histogram of Oriented Optical Flow and Object Contour Analysis. For the Computational Intelligence stage we use a Self-Organizing Map (SOM) optimized through Learning Vector Quantization (LVQ). The approach is then applied for classification of human actions in many real situations. Testing against a database with different kinds of human actions, we show the usefulness and robustness of this method, comparing it to other proposals in the literature.
Keywords :
computer vision; image classification; image motion analysis; image sequences; learning (artificial intelligence); object recognition; self-organising feature maps; vector quantisation; LVQ; SOM; artificial neural networks; computational intelligence; computer vision; human action classification; human action recognition; image sequence; learning vector quantization; motion analysis technique; object contour analysis; oriented optical flow histogram; selforganizing map; Computational intelligence; Computer vision; Feature extraction; Histograms; Humans; Image sequences; Optical imaging; histogram of oriented optical flow; human action recognition; object contour analysis; self-organizing map (SOM);
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Neural Networks (IJCNN), The 2012 International Joint Conference on
Conference_Location :
Brisbane, QLD
ISSN :
2161-4393
Print_ISBN :
978-1-4673-1488-6
Electronic_ISBN :
2161-4393
Type :
conf
DOI :
10.1109/IJCNN.2012.6252612
Filename :
6252612
Link To Document :
بازگشت